Cloud Digital Repository Automation Matthew Brockman, Chris Hill Multimedia, Hypertext, and Information Access, CS-4624 Edward A. Fox Virginia Tech, Blacksburg VA 24061 5/2/2018

Outline Continuous Integration and Continuous Deployment Overview Tools Used Completed Pipeline Analyzing Our Cost Final Thoughts Final Work before Submission to VTechWorks

CI/CD Continuous Integration and Continuous Deployment Software engineering practice Used in Agile environments One of the main, and most difficult components of the DevOps process

Fedora4 Our Task: CICD for Fedora4 Any Development Change prompts a build/test/deployment Manage bugs and feature changes more efficiently Always has a updated and working version of the application Overhead for deployment and manual unit testing becomes automated Helps the Fedora Dev team

Our AWS Stack AWS Elastic Beanstalk CodePipeline CodeBuild Simple Cloud Storage Service (S3) Elastic Beanstalk CloudWatch AWS Elastic Beanstalk

Finished Pipeline Final View: Awaiting Review No Deployment Email based Approval Email subscription can be setup/modified easily

Finished Pipeline

Finished Pipeline Final View: Finished and all Green Each Stage Completed Pipeline triggered from source change

Total Costs from the Pipeline Minimal Costs Amazon S3 $0.023 per first 50 TB CloudWatch Free / Logging Service Possible Larger Costs CodeBuild Build Minute Based Elastic Beanstalk (ie EC2)

Total Costs EC2 instance At 100% utilization Hosting Opt 1 $35 a month Upfront cost for a dedicated server Hosting Opt 2 $103 per year CodeBuild daily builds $21.90 per year CodePipeline pipeline $1.00 per pipeline Roughly $126 to host and build per year

Final Thoughts Groundwork Pipeline Minimal changes needed to handoff permissions and access tothe Pipeline with transitioning to client Room for more stages, configurations, and plug-ins Providing an AWS starting point for the Fedora4 team Modest cost for an automatic process

Finishing Touches Document of AWS roles/users so that Fedora devs can follow and modify the Pipeline Cleaning up unused AWS resources (Storage/Invalid Builds) Screenshots/Commenting our Configs Finalizing a cost spreadsheet Synthesizing our Final Report Including Cost spreadsheet, AWS accounts, permissions, and screenshots we have been using for the AWS stack

Acknowledgements Client: Yinlin Chen ylchen@vt.edu (540-231-0271) Fedora4 Development Team and DuraSpace Organization

References AWS CodePipeline/CodeBuild/CodeDeploy/ElasticBeanstalk : https://docs.aws.amazon.com/codepipeline/latest/userguide/codepipeline-user.pdf https://aws.amazon.com/documentation/elastic-beanstalk/ https://aws.amazon.com/documentation/codebuild/ https://aws.amazon.com/documentation/s3/ https://aws.amazon.com/codebuild/pricing/ https://aws.amazon.com/ec2/pricing/ Monthly Calculator: https://calculator.s3.amazonaws.com/index.html Images (in order): https://loadimpact.com/integrations/performance-testing-aws-codepipeline https://docs.aws.amazon.com/codepipeline/latest/userguide/welcome.html https://aws.amazon.com/blogs/security/how-to-audit-cross-account-roles-using-aws-cloudtrail-and-amazon-cloudwatch-events/ https://www.flickr.com/photos/pictures-of-money/17309480255 https://cloudacademy.com/blog/elastic-beanstalk-lab/